Agent as Server

API documentation says that the agent can act as a server.
How many incoming requests can this server handle?
Did requests are set in the line to proceed? I mean - did I must re-send a request?

We have some sample code that might help you.

I know that. My question was about connections to the server, mau be not about code for Agent to act as server.

Rate limits on inbound HTTP are a little variable (there are multiple backends which load is distributed over and the rate limit is not shared between these to reduce coupling). You’re generally fine with around 20 requests per second though - assuming you’re completing them quickly.

When you hit the limit you’ll see 429’s returned; these requests won’t make it to the http.onrequest handler. You can retry the 429’s later, obviously.

Does that help?

Thanks, Hugo!
Yes - that help.